darmanin-matt/smnli_mt

---
license: cc-by-sa-4.0
task_categories:
- text-classification
language:
- mt
tags:
- nlp
- nli
- rte
- maltese
- malti
size_categories:
- 100K<n<1M
configs:
- config_name: snli
data_files: "snli_mt.csv"
sep: ";"
- config_name: mnli
data_files: "mnli_mt.csv"
sep: ";"
---

# Dataset Card for the SMNLI-MT

The SMNLI-MT datasets are machine-translated versions of the Stanford NLI and MultiNLI datasets in Maltese.

## Dataset Details

### Dataset Description

The datasets were translated using the Google Cloud Translate as part of the initial exploration of NLI in the Maltese language.

- **Curated by:** Matthew Darmanin
- **Language(s) (NLP):** Maltese
- **License:** CC 4.0

## Dataset Structure

The datasets are in the form of CSV files, delimited by semi-colons (;) and encoded in UTF-8.

Each dataset contains the following columns:
- **id** - the index of the sentence pair
- **premise** - the premise sentence
- **hypothesis** - the hypothesis sentence
- **gold_label** - the consensus label assigned in the original datasets; either _entailment_, _contradiction_ or _neutral_

## Dataset Card Contact

E-mail: <[e-mail]>

原始信息汇总

数据集卡片 for SMNLI-MT

SMNLI-MT 数据集是斯坦福自然语言推理（NLI）和多NLI数据集的机器翻译版本，翻译成马耳他语。

数据集详情

数据集描述

这些数据集是使用谷歌云翻译作为马耳他语NLI初步探索的一部分进行翻译的。

• 由以下人员策划： Matthew Darmanin
• 语言（NLP）： 马耳他语
• 许可证： CC 4.0

数据集结构

数据集以CSV文件形式存在，使用分号（;）分隔，编码为UTF-8。

每个数据集包含以下列：

• id - 句子对的索引
• premise - 前提句子
• hypothesis - 假设句子
• gold_label - 原始数据集中分配的共识标签；可以是_entailment_、contradiction_或_neutral

数据集卡片联系

电子邮件：[e-mail]